EXPERIMENT STATION BULLETINS. 453 



soil at five different moistures 5 x 12=60 flasks to be weighed out, steri- 

 lized and inoculated separately. This amount allowed duj)licate deter- 

 minations at five different times and one duplicate set of blanks. The 

 peptone, 0.5 g. per .50 <;•. of oven-dry soil, was mixed with the soil before 

 weighing out the individual sample, also the water in all cases where the 

 soils were not too moist. The samples, each corresponding to 50 g. 

 oven-dry soil, were placed into liter flasks, sterilized at 15 pounds pressure 

 for 20 minutes, and stored in a room with a constant temperature of 18^ 

 until all the flasks for one set had been sterilized. They were then in- 

 oculated with l^ cc. of a 24 to 48 hours old culture of B. mycoides which 

 had been transplanted in regular intervals in peptone solution. The 

 culture was distributed in the soil by shaking as well as could be done; 

 Avhile it was not difficult 1o accomplish this with the sands and part of 

 the soils, others contained enough loam to cake and the distribution 

 seemed rather irregular. All the cultures were shaken again after one 

 day's growth. 



The ammonia in these cultures was determined after two, four, six, 

 ten and twenty days by distilling it over with water and MgO into n/10 

 acid. Bacterial counts could not be made in this series, for lack of 

 time, since the 30 distillations (three soils at five different moistures in 

 duplicates) occupied practically a whole day. In some of the later ex- 

 periments bacterial numbers and ammonia have been determined at the 

 same time. 



Data. — Table 1 gives all results of this first experiment. They are 

 given in cc. n/10 ammonia per 50 g. of dry soil, after subtraction of the 

 ammonia in the blanks. Instead of recording each of the duplicates as 

 it was found, it seemed preferable to give the average of the two and the 

 deviation. Thus, if the reading of the table is 7.2 cc. and the error 

 +0.10, it means that the two determinations gave 7.1 and 7..3 cc. This 

 way of recording allows a more ready comparison of the data and of the 

 deviations by themselves. 



This entire experiment with nearly 1,100 single determinations has 

 been carried out by Mr. L. R. Himmelberger, to whom I am much in- 

 debted for this mass of data. 



